• Quality bonus update

    Q: When is a quality bonus not for quality?

    A: When everybody gets it.

    This is what’s happening with the Medicare Advantage quality bonus expansion I posted about this morning.  You may recall that CMS announced on Veterans’ Day that bonus payments originally targeted to high-quality plans (those enrolling about 24% of beneficiaries) would be expanded to plans with lower quality scores too.  Under the proposed expansion, at least 84% of beneficiaries will now be in plans qualifying for bonuses—more than tripling the size of the bonus program.

    In this morning’s post I questioned the cost estimate for this expansion and suggested that the motive behind it may be to gain political support from America’s Health Insurance Plans (AHIP).  In response to that post, a reader forwarded to us details about the proposed expansion which I had been unable to obtain from the CMS press office or website.   The current proposal does not duplicate some of the more expensive features of the old program (e.g. double-bonuses for certain plans) so it will be a lot cheaper than my first estimate.  I have redone my cost estimates and the price tag now comes in at $16.1 billion, still more than 3 times the official estimate projected over ten years.*

    I think, had the details been available yesterday, I would have written essentially the same post this morning with the lower number.  A “quality bonus” that’s given to almost all plans is really just a back-door rate increase and $16 billion is a lot of money to give away.  Furthermore, the regulations are not yet final and, as you can see from how these estimates have changed, seemingly small details of the program can trigger large increases in spending.  We will be watching as this continues to develop, but the potential for hidden mischief is worrisome.

    *CMS may object, saying that the new bonuses are only in force for three years under demonstration authority.  I don’t believe they will go away once put in place, so I think the customary ten-year budget window is appropriate.

    • I think I disagree. The proposed MA quality indicators exceed existing HEDIS (PQRI, et. al) indicators in their completeness. Two of the proposed measures are more ‘outcome’ focused, as opposed to ‘output’ focused. The last two of “staying healthy” are amazingly difficult to quantify,
      “-The percentage of members who improved or maintained their physical health after two years
      -The percentage of members who improved or maintained their mental health after two years”

      Those are outcome measures unlike anything I’ve seen done with big programs and I’m very curious to see how they’re defined. HEDIS doesn’t do anything like it currently (http://www.ncqa.org/Portals/0/HEDISQM/HEDIS%202011/HEDIS%202011%20Measures.pdf).

      Indeed, as someone who works in quality, I find the prospect of getting people together in a room to define those numerators and denominators somewhat terrifying.

      Brian Biles and Grace Arnold of George Washington University based their analysis on plans that were not audited against those metrics. If I’m wrong on that point, I welcome correction.

      It’s almost an integrity issue I take with the paper as their appendix 1 is acting as if a perfect crosswalk can be made with the new indicators and previous indicators and I take strong issue with that assumption.

      Maybe a hyperbolic comparison but it is sort of like assuming a top rating from Zagat’s is the same as a top rating from Michelin.

    • more on this, confirmed here:


      these indicators do not crosswalk with these:

      so it is not correct to say, “84% of beneficiaries will now be in plans qualifying for bonuses,” because no plans have been held to this standard.

    • Thanks for bringing up this issue, but am I missing something? The CMS factsheet gives a distribution of plans and says, “This star rating distribution is the same as the star rating distribution for purposes of the quality bonus determination under the demonstration.” So CMS seems to think the 84% figure reflects the ratings they plan to use.

    • I think it reflects the 2011 distribution, but the 2012 distribution will be with the new criteria and, presumably, a different distribution as a result, no?

    • I don’t think so. I went back and checked some data I downloaded from CMS earlier and they have ratings for these two elements from past years. I don’t see any reason why the 2012 distribution would differ very much.

    • hmm…I wonder if those are some sort of aggregate measures. Maybe they use 4 or 5 of the HEDIS measures to come up with that value.

      Kind of disappointing, actually, as I was excited about some of those more audacious measures.

      Do you have a link for the historical data on those? Ok to send offline.

    • Sorry to be slow responding– I just saw your latest comment. It was a while ago, but I believe I downloaded these data from Medicare.gov. Go to https://www.medicare.gov/find-a-plan/questions/home.aspx and click the link at the lower right called “Download the Medicare Health Plan Compare Database”. You will have to choose which database you want, so pick “Plans – Quality Data”. You can get an Access database or a comma-delimited flat file. Most of the ratings are national, so the files aren’t very big.